High inequality in the heart of Africa: the case of Rwanda

Transcription

1 High inequality in the heart of Africa: the case of Rwanda Andy McKay WIDER Inequality Conference, 5-6 September 2014

2 Inequality in Eastern Africa Some summary Gini coefficients: Country Year Gini coefficient Burundi Kenya Tanzania Uganda Rwanda Ethiopia Malawi Mozambique Rwanda Data from WIID, latest issue

3 Inequality in Eastern Africa Rwanda has highest Gini coefficient in EAC And high compared to most countries in bigger region Though higher inequality in some Southern African countries (also Central African Republic) And a very low estimate of inequality from 1984 though provenance not clear

4 Structure Introduction to Rwanda Consumption inequality Looking at income data Land? Economic activity and income sources Extending back to 1990? What can we say?

5 The case of Rwanda Small country with highest population density in Sub-Saharan Africa Dominantly agricultural economy History of conflict culminating in 1994 genocide Good quality survey data from 2000 onwards: source of available estimates of inequality (and poverty)

6 Rwanda: recent economic performance Impressive growth rate of consumption of 4.4% in 2005/6 to 2010/11 (also national accounts) And good recent record of poverty reduction: poverty fell from 58.9% in 2000/01 to 56.7% in 2005/06 to 44.9% in 2010/11 These estimates are based on adjusted real household consumption per adult

7

8 Headline consumption inequality numbers (national) Table shows different indices for the three years; High levels All indices increase 2000/1-2005/6 and fall between 2005/6 and 2010/11 (as GICs) p90/p10 GE(0) GE(1) Gini 2000/ ( ) ( ) 2005/ ( ) ( ) 2010/ ( ) ( )

9 Disaggregated Gini coefficients By stratum, province. Big urban-rural gap, Kigali vs. rest 2000/ / /11 by stratum City of Kigali Other urban Rural by province Kigali City Southern Province Western Province Northern Province Eastern Province National

10 Theil index decomposition Between stratum inequality accounts for 25-30% of total; yet high urban inequality 2000/ / /11 by stratum City of Kigali Other urban Rural proportion of b/w stratum variation by province Kigali City Southern Province Western Province Northern Province Eastern Province proportion of b/w province variation proportion of b/w district variation

11 Income inequality Survey data enables computation of different household income components Important because captures livelihoods Quality of income data seems adequate; underestimation, but decreasing with time Positive correlation association even removing common elements

12 Some income inequality numbers Income inequality numbers higher (no surprise) Gini 2000/ / /11 Theil 2000/ / /11 National by stratum Kigali Other urban Rural proportion of between stratum variation by province City of Kigali Southern Western Northern Eastern proportion of between province variation

13 Income inequality Income inequality numbers show no clear trend But do show the same urban-rural gap, and between Kigali and other provinces Consumption inequality data more reliable, but pattern largely confirmed here

14 How important is land inequality? Survey has self reported information on plot areas but plots can be very different Land inequality high (e.g. percentile ratios); and many have very small areas Gini similar trend to consumption inequality; and % with small area linked to quintile % of farming households with less than p90/p10 GE(1) Gini 0.2Ha 0.5Ha 2000/ / /

15 Economic activities Type of activity households able to undertake likely to be strong correlate of inequality Income source data to define economic activity groups: main income source or different diversified patterns Agriculture dominates, except in 5 th quintile; non-farm wage work and business much more important in 5 th quintile Big increase in agriculture plus farm wage in 2010/11

16 Economic activities: distribution Increase in diversification even in Q1; and small increase in nonfarm wage work Livelihood status 2000/ / /11 % of Q1 % of Q5 % of Q1 % of Q5 % of Q1 % of Q5 Mainly agriculture Mainly farm wage Mainly nonfarm wage Mainly business Mainly transfers/rent Agric/farm wage Agric/nonfarm wage Agric/transfers Agric/business Other combinations All

17 Economic activities: inequality High inequality in non-farm activities and transfer recipients; much smaller in agriculture based activities Livelihood status Gini 2000/ / /11 Mainly agriculture Mainly farm wage Mainly nonfarm wage Mainly business Mainly transfers/rent Agric/farm wage Agric/nonfarm wage Agric/transfers Agric/business Other combinations All

18 Income source by quintile Table shows income shares from 2010/11; agriculture high except Q5; nonfarm wage important in all, esp. Q5; farm wage in Q1 and public transfers consumption quintile agriculture farm wage nonfarm wage nonfarm nonwage public transfers private transfers Lowest 46.0% 17.3% 14.3% 9.5% 8.2% 4.6% Second 56.0% 10.5% 14.2% 10.8% 2.9% 5.5% Third 55.4% 6.6% 13.6% 16.5% 2.6% 5.3% Fourth 46.3% 4.0% 16.3% 24.7% 3.6% 5.1% Highest 12.1% 0.7% 38.2% 40.4% 1.6% 7.0% All 27.8% 3.8% 28.7% 31.0% 2.6% 6.3%

19 Income source decomposition Decomposition of Gini coefficient by income source (here 2005/6, others similar) Agriculture less unequal, lower correlation: smaller contribution to total inequality Nonfarm activities contribute to inequality Source Sk Gk Rk Share Agriculture Farm wage Nonfarm wage Nonfarm nonwage Public transfers Private transfers

20 Income analysis: summary Significantly lower inequality among those in agriculture Impact of land inequality not just seen here Increased diversification over period, including for poorest Fewer nonfarm activities in lower quintiles, but also nature of activity very different Some suggestive evidence that public transfers may reach poor groups

21 Extending the analysis back to 1990 Household surveys started from 2000/01 But were a long series of agricultural surveys in Rwanda from 1980s on, one in 1990 collecting information on income and food expenses Methodology is different: but seek to compute measures of income and food expenditure as comparable as possible Rural areas

22 Inequality back to 1990 First look at trends in per capita income and food consumption Income inequality suggests higher inequality from 2000 on, but food consumption does not p90/p10 GE(1) Gini income / / / food consumption / / /

23 Inequality back to 1990 (cont) Food consumption might be more accurately measured but inequality in nonfood consumption may still have increased Income inequality suggests higher inequality from 2000 on, but food consumption does not

24 Inequality back to 1990 If we look at disaggregated data (old province structure), food consumption inequality measures much more stable Province Theil Index Gini / / / /06 Butare Byumba Cyangugu Gikongoro Gisenyi Gitarama Kibungo Kibuye Kigali Ngali Ruhengeri

25 Total Changing income structures back to 1990 Table shows distribution by same groups Share of mainly agriculture very similar except same diversification in 2010/11 Slow growth in wage and business, esp. in 2010/ / / /11 Mainly agriculture Mainly farm wage Mainly nonfarm wage Mainly business Mainly transfers/rent Agric/farm wage Agric/nonfarm wage Agric/transfers Agric/business Other combinations

26 Income structure decomposition back to 1990 Source Sk Gk Rk Share 1990 Agriculture Farm wage Nonfarm wage Nonfarm nonwage Transfers /01 Agriculture Farm wage Nonfarm wage Nonfarm nonwage Transfers /06 Agriculture Farm wage Nonfarm wage Nonfarm nonwage Transfers /11 Agriculture Farm wage Nonfarm wage Nonfarm nonwage Transfers

27 Income structure decomposition back to 1990 Similar to national patterns: agriculture and farm wage generally contribute less to rural food consumption inequality Nonfarm activities tend to increase it Transfers may reduce it (but small)

28 What can we say? Lots of numbers, and varying degree of confidence in different sources of data. Some interim conclusions 1. Inequality in Rwanda is high: not that much less than Brazil now. Consumption, income and land all tell the same story. 2. Rural inequality though may not have changed that much in 20 years

29 What can we say? (2) 3. Recent consumption and land data suggests a reduction 2005/6-2010/11, but is that real and sustainable? Shape of 2005/6-2010/11 GIC partly explained by agricultural harvests, though transfers and less land inequality may play a part 4. There is a very big urban-rural and Kigalirest gap, and very high urban inequality. Migration; umudugudu policy; Kigali development etc.?

30 What can we say? (3) 5. Nonfarm activities a big driver of inequality, including within rural; poorer have less good access but also get access to poorer opportunities 6. In 2010/11 more diversification, though may be in part good harvests that year and labour demand.

31 What can we say? (3) 5. Nonfarm activities a big driver of inequality, including within rural; poorer have less good access but also get access to poorer opportunities 6. In 2010/11 more diversification, though may be in part good harvests that year and labour demand.

32 Thank you!